You are on page 1of 4

c.pelling@bbk.ac.

uk

B.A. Introduction to Logic 2012-13


Lecture 6: Propositional Logic IV
Recap
Rules of inference: MP: Given a conditional formula on one line of proof and its antecedent on another line, its consequent may be inferred. The line numbers of each must be cited together with MP. The dependency-numbers of the new line consist of the dependency-numbers of both cited lines. A: Any formula may be assumed on any line of proof. The line must be annotated A for assumption. The dependency-number of the assumed formula is identical with the line number of the line on which it is assumed. CP: Assume the antecedent. Derive the consequent. Enter the conditional on a line along with CP. The line numbers are both those where the antecedent is assumed and those where the consequent is derived. Discharge the dependency-number of the antecedent. Pool the remaining dependency-numbers to complete the line.

A constraint on the correct use of CP


Note the following constraint on the correct use of CP: in deriving the consequent of the conditional youre trying to prove, you must actually use your assumption of the antecedent (together, if necessary, with other formulas that are available). In some cases, its not immediately obvious how to meet this constraint, because youre faced with the following problem. Although you have the consequent of the relevant conditional on a line in your proof, its not clear how you can derive it using your assumption of the antecedent. Example: Suppose youre trying to prove the sequent: Q:PQ We start in the natural way: {1} {2} 1. 2. Q P Premise A

Now we want to derive Q. But we have a problem. Because although we already have that formula on line 1, its not clear how we can derive it using the assumption we made on line 2. The solution is to make clever use of the rules &I and &E, as follows: {1,2} {1,2} 3. 4. P&Q Q 1,2 &I 3 &E 1

c.pelling@bbk.ac.uk Although the formula on line 4 is the same as the formula on line 1, they have different dependencynumbers: in particular, the dependency-numbers for line 4 include line 2, where we assumed P. This means that strictly speaking, the assumption has been used to derive the formula on line 4, and so we can go on to use CP to get our conclusion. {1} 5. PQ 2,4 CP

Another example
Suppose were trying to prove the sequent: P, P Q : R Q We do it as follows, again making clever use of &I and &E as we go: {1} {2} {3} {1,2} {1,2,3} {1,2,3} {1,2} 1. 2. 3. 4. 5. 6. 7. P PQ R Q R&Q Q RQ Premise Premise A 1,2 MP 3,4 &I 5 &E 3,6 CP

Note the difference between the dependency-numbers on line 4 and those on line 6. Since those on line 6 include the line number of our assumption, were entitled to use CP to derive our conclusion.

Theorems
Some valid sequents in PL are peculiar in that they have no premises. These sequents are known as theorems. Question: How do you prove a theorem, if you havent got any premises to work with? Answer: Theorems can be proven purely by using assumptions. Example: Suppose we want to prove the theorem: : P (Q (P & Q)) The conclusion is a conditional, so we start by assuming the antecedent: {1} 1. P A

Now we want to derive the consequent, which is itself a conditional: Q (P & Q) So the next step is to assume the antecedent of that conditional: 2

c.pelling@bbk.ac.uk {2) 2. Q A

Now were looking to derive P &Q. We do this in the obvious way, using &I: {1,2} 3. P &Q 1,2 &I

Now we use CP to derive Q (P & Q), discharging our assumption in line 2 in the process: {1} 4. Q (P & Q) 2,3 CP

Finally we use CP a second time to derive our conclusion, discharging our assumption in line 1 in the process: --5. P (Q (P & Q)) 1,4 CP

Biconditionals
Consider the sentence Juventus won the title if and only if they beat AC Milan. If true, this seems to leave open two possibilities: 1) Juventus won the title and they beat AC Milan 2) Juventus didnt win the title and they didnt beat AC Milan But two other possibilities are ruled out: 3) Juventus won the title but they didnt beat AC Milan 4) Juventus didnt win the title but they did beat AC Milan If this is the correct analysis of the original biconditional sentence, its natural to think of it as equivalent to the conjunction of two ordinary conditionals: If Juventus won the title then they beat AC Milan (this rules out possibility 3) If Juventus beat AC Milan then they won the title (this rules out possibility 4)

On this view, the original biconditional sentence is equivalent to if Juventus won the title then they beat AC Milan, and if they beat AC Milan then they won the title. Parallel remarks apply to the relation between the PL connectives and : for example, the formula P Q can be considered equivalent to the formula (P Q) & (Q P). This will become clearer when we look at the inference rules for , which is our next task.

The rules I and E


Example: Suppose we want to prove this sequent: PQ:QP We start by entering the premise: 3

c.pelling@bbk.ac.uk {1} 1. PQ Premise

For our next step, well need to use the elimination rule for :

E: Given a biconditional on a line we may rewrite that formula as the conjunction of the relevant
pair of conditionals on the next line of proof. The new line is annotated with the line number of the old line and takes as dependency-numbers all and only those of the old line. Using this rule, we can write: {1} 2. (P Q) & (Q P) 1 E

Our next move will be to break this down into two separate conditionals, using &E: {1} {1} 3. 4. PQ QP 2 &E 2 &E

Now we can derive our conclusion, by using the introduction rule for :

I: Given a pair of conditionals on two lines such that the antecedent of the first is the consequent of the second and the consequent of the first is the antecedent of the second, you may write between
the two formulas, antecedent and consequent, on a new line. The new line is annotated with the line numbers of both lines used and I. The dependency-numbers of the new line are all of those of both lines used. Using this rule, we can write: {1} 5. QP 3,4 I

Reading
Tomassi, P. Logic. Chapter 2, VIII XI.

Exercises
Exercises 2.6 & 2.7

http://fundraise.unicef.org.uk/MyPage/Charlie-KP-Marathon

You might also like